Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

In non-standard analysis, assuming the continuum hypothesis, the field of hyperreals $\mathbb{R}^*$ is a field extension of $\mathbb{R}$. What can you say about this field extension?

Is it algebraic? Probably not, right? Transcendental? Normal? Finitely generated? Separable?

For instance, I was thinking: Would it be enough to adjoin an infinitesimal to $\mathbb{R}$ and get $\mathbb{R}^*$? The axioms of hyperreals in Keisler's Foundations of Infinitesimal Calculus seem to suggest so, but I'm not sure.

I don't know how to approach this question, since infinitesimals and such things don't "result" from polynomials (like, say, complex numbers do).

Is $\mathbb{R}^*$ maybe isomorphic to the field of fractions of some polynomial ring? (This occured to me after thinking about $\mathbb{R}(\epsilon)$ and such things, where $\epsilon$ is an infinitesimal).

share|cite|improve this question
I just realised it's quite easy to show it isn't algebraic. But what about the rest? – FPP Nov 10 '12 at 22:24
There actually exists no single field of hyperreals. Different fields of hyperreals may fail to be isomorphic. – Michael Greinecker Nov 10 '12 at 22:24
Let's assume the continuum hypothesis, then. If I remember correctly, that implies that all are isomorphic, right? – FPP Nov 10 '12 at 22:25
@FFP Yes.${}{}$ – Michael Greinecker Nov 10 '12 at 22:26
Well, exactly what should $\epsilon$ satisfy? How is (an) $\Bbb R^*$ constructed? – Berci Nov 10 '12 at 23:18

There isn't any trouble constructing $\mathbb{R}(\epsilon)$ as a formally real field that is isomorphic (as a field) to $\mathbb{R}(x)$, and with $\epsilon$ a positive infinitesimal. (it may be easier to see the ordering by writing a rational function as a formal Laurent series in $\epsilon$ about 0)

It's easy to see that this is not a hyperreal field: it only has finite powers of $\epsilon$. In particular, if there were a positive transfinite positive integer $H$, then there would be an element $\epsilon^H$ that is smaller than every power of $\epsilon$ appearing in $\mathbb{R}(\epsilon)$.

($\mathbb{R}(\epsilon)$ also fails to have a square root of $\epsilon$. This can be fixed by passing to its real closure, but that would still fail the above property)

If you tried adjoining a second infinitesimal, a similar argument proves it can't be hyperreal.

share|cite|improve this answer

Briefly, one way to approach this question would be to try to construct alternative models of the hyperreals via the Compactness Theorem that either satisfy or fail to satisfy the properties you want. The Enderton text, A Mathematical Introduction to Logic, uses this kind of construction and has a very rigorous treatment of nonstandard analysis that you can use as a guide.

Since I only have 44 minutes left on this question, it could easily contain errors so I'm retagging your question to include logic and set theory so it can be properly checked.

share|cite|improve this answer

I don't think it would be enough to adjoin one infinitestimal. To construct the hyperreals via an ultrapower requires the Axiom of Choice, but you could certainly adjoin one infinitesimal without Choice.

share|cite|improve this answer

A small correction to the formulation of the original question: the continuum hypothesis is not needed to construct a hyperreal field extension of the reals. The alternative simpler field extension you are proposing is related to the Levi-Civita field, which was of intense interest to Abraham Robinson. However, it does not have the crucial properties that make the hyperreals useful in analysis.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.