MathOverflow is a question and answer site for professional mathematicians. 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

For $m,n,k < \omega$, consider the equation

$X \to (\omega \times k)^{m}_{n}$

What is the smallest $X$ known to satisfy it?

Baumgartner-Hajnal theorem gives a satisfactory answer for $m=2$, but it fails badly for $m>2$, and actually I hardly managed to find anything in the literature concerning $m>2$. The only bound I know is the one given by Erdos-Rado - $\beth_{\omega}$. So $\aleph_{\omega}$ works assuming $GCH$. Can it be proved to work in $ZFC$? Any improvements and/or references are very welcome.

Also I would be happy to hear any results for the same equation with $\omega$ replaced by an arbitrary ordinal $\lambda$.

share|cite|improve this question

Here is what I know about this (and mostly learned from my student Thilo Weinert):

It is known that $\mathbb Q\not\rightarrow(\omega+1,4)^3$. I.e., there is a coloring of the unordered triples of rationals with two colors $0$ and $1$ such that no set of rationals of ordertype $\omega+1$ is homogeneous of color $0$ and no set of size 4 is homogeneous of color 1.

On the other hand, Milner and Prikry showed $\omega_1\rightarrow(\omega\cdot 2,4)^3$ and Jones proved $\omega_1\to (\omega+m,n)^3$ for all $n,m\in\omega$. Milner and Prikry conjectured that $\omega_1\rightarrow(\alpha,n)^3$ holds for all $n\in\omega$ and all countable $\alpha$.

Note that all the positive results are unsymmetric, i.e., they only give finite homogeneous sets in one of the colors.

Also, there is a definable (continuous, actually) coloring of the unordered triples of $2^\omega$ (ordered lexicographically) with two colors such that every infinite homogeneous set has ordertype either $\omega$ or $\omega^*$.

Finally, let us consider this coloring: for each $\alpha\lneq\omega_1$ fix a wellordering $\leq_\alpha$ of $\alpha$ of ordertype $\leq\omega$.
Given three ordinals $\alpha\lneq\beta\lneq\gamma\lneq\omega_1$, let $c(\alpha,\beta,\gamma)=0$ if $\leq_\alpha$ agrees with the usual ordering of ordinals on $\{\alpha,\beta\}$ and otherwise let $c(\alpha,\beta,\gamma)=1$.

I think there is no homogeneous set of color $0$ of ordertype $\omega+2$ and no homogeneous set of color $1$ of ordertype $\omega+1$. This would show $\omega_1\not\rightarrow(\omega+2)^3_2$.

share|cite|improve this answer
Thanks for your answer, Stefan. However, it does not quite address my concern. I mean, these are explications of saying that for $m>2, k>1$ we can not take $\omega_1$ as $X$. But then what can we take? How about $\omega_2$? – Artem Chernikov Jul 29 '11 at 13:00
I thought for a while that it might be possible to show $\omega_n\not\rightarrow(\omega+n+1)_2^{n+2}$, but the proof I had in mind doesn't seem to work. Hence I removed my conjecture about this from my answer. – Stefan Geschke Jul 30 '11 at 21:22
I see. But do you still expect it to be true? Also, now I know that what I actually want to solve is $X \to (\omega+k)^{m}_{n}$. Aren't there any positive results at all with bounds better than the one given by Erdos-Rado, that is $\beth_{\omega}$? – Artem Chernikov Jul 30 '11 at 21:42

I think that the general conjecture $\omega_1\to(\alpha,n)^3$ goes back to Erdos and Rado.

share|cite|improve this answer
Thanks, Peter, I was not aware of that, but I have not personally looked at the Milner-Prikry paper either, I have to confess. – Stefan Geschke Jul 29 '11 at 13:48

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.