MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

I come up with the following set-theoretic question that has the flavor of Maximal Almost Disjoijnt (M.A.D.) families, although it is a bit different than the usual setting:

Let $\kappa$ be an uncountable cardinal. I want to find $\lambda$ such that:

1) there is a family of functions $f_\alpha:\lambda \rightarrow \kappa$, for all $\alpha< \lambda$,

2) for every $\alpha,\beta<\lambda$, $f_\alpha\cap f_\beta$ is finite (or empty), i.e. there are only finitely many $x\in \lambda$ such that $f_\alpha(x)=f_\beta(x)$, and

3) $\lambda$ is the greatest possible such cardinal.

Alternatively, we can see this family as one function $f:\lambda\times\lambda \rightarrow \kappa$ and the difference is that $\lambda$ controls both the domain of the functions as well as the number of them.

Now, it is trivial to construct a family of size $\kappa$ that satisfies (1) and (2) and it is not too hard to prove that it can not be done for $\lambda=\kappa^{++}$. So we are left we only two possibilities. Either $\lambda=\kappa$ is the best possible, or $\lambda=\kappa^{+}$ is the best possible.

Under the additional assumption $\kappa^\omega=\kappa$, we can prove that $\lambda=\kappa$ is the best possible.

My question: Does anyone know any references that relate to this question?

share|cite|improve this question
Some of us instinctively read "M.A.D." as either "median absolute deviation" or "mean absolute deviation". (But ignore this silly comment.) – Michael Hardy May 20 '11 at 2:35
If MAD was "median absolute deviation", it would not be a MAD family. Right? – Ioannis Souldatos May 20 '11 at 19:05

The article:

"Mad families and their neighbors" (2007) by Andreas Blass, Tapani Hyttinen, Yi Zhang

(pdf available here:

talks about several related notions of MAD, and might be a decent starting point.

This is an intriguing construction. In passing I would like to point out that $\kappa^\omega = \kappa \implies \lambda = \kappa$, says that under $GCH$ (so in $L$ for example) the only places the assertion can fail is at a singular cardinal of countable cofinality. As such you might want to have a look at what happens at $\omega_\omega$ under $CH$ (where you have $\forall n > 0 ({\omega_n}^{\omega} = \omega_n)$.) Or more to the point, what happens when you have a sequence of cardinals $\kappa_n : n\in \omega$, such that for each $n\in\omega$, $\lambda_n = \kappa_n$, and $\kappa = \sup\{ \kappa_n: n\in\omega \}$.

share|cite|improve this answer
Given that $\kappa^\omega=\kappa\Rightarrow \lambda=\kappa$, my first naive idea is that $\lambda$ is the minimum of $\kappa^+$ and $\kappa^\omega$. In particular, can we have $\kappa^\omega>\kappa$ and still $\lambda=\kappa$? I am not sure how difficult these questions are, but I do not have any examples when $\lambda=\kappa^+$, except for $\kappa=\aleph_0$. The argument (by Hjorth) uses Descriptive Set Theory and I wanted to see if anyone else knows anything about it. – Ioannis Souldatos May 20 '11 at 19:03
Well, what I was alluding too was trying to glue an \omega sequence of these families together, and see if the result is such a family. – Michael Blackmon Jul 29 '11 at 16:11

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.