Question goes as follows : "What can be written instead of the supremum axioms, to get the same structure in $\mathbb{R}$." Now for the supremum axioms I have : If $K_a \neq Ø \Rightarrow \exists$ a smallest upper bound $ \exists min$ $K_a \iff \exists \alpha \in K_a$ $\forall k \in K_a : \alpha \leq k$

$ \alpha = sup(A) $ ergo the least upper bound.

  1. $ \alpha$ is upper bound, $ \forall a \in A : a \leq \alpha$
  2. $ \alpha$ is the smallest/least upper bound, $ \forall \epsilon \gt 0 $ $\exists a \in A : a \gt \alpha - \epsilon$

Now my question is what can I write instead of these axioms. I was thinking some kind of isomorphism to preserve the structure, but I am not sure.


1 Answer 1


If you're hoping for an alternate axiom which is "purely algebraic" - that is, avoids talking about arbitrary sets of reals, and is closer to (say) the field axioms - then there is a sense in which no such axiom exists: specifically, any first-order axioms will fail to capture the real numbers, in the following sense. If $T$ is a set of first-order sentences which are true about the structure $(\mathbb{R}; +, \times, <)$, then

  • There will be models of $T$ which are "too big", e.g. have cardinality much greater than that of $\mathbb{R}$. In particular, there will be models which are non-Archimedean (an ordered field is Archimedean if for any $x>0$, there is some natural number $n$ such that $1+1+...+1$ ($n$ times) is $>x$).

  • There will be models of $T$ which are "too small," e.g. are countable. Some of these will be non-Archimedean, but we'll also have proper substructures of the reals satisfying $T$.

This is due to the compactness and Lowenheim-Skolem theorems of first-order logic, respectively.

That said, we can certainly avoid sups and infs! $\mathbb{R}$ is a maximal Archimedean ordered field; and any two maximal Archimedean ordered fields are isomorphic. Note, however, that this is non-first-order in two ways:

  • We have to say that $\mathbb{R}$ is Archimedean.

  • We have to say that any proper ordered field extension of $\mathbb{R}$ is not Archimedean.

There is a precise sense in which this second sentence is "even less first-order" than the first; but that's getting a bit off-track. If you're interested in these sorts of things, you might look at abstract model theory.


You must log in to answer this question.

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