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

Several days ago a friend asked me the following:

We know that in $\mathcal P(\mathbb N)$ we can find a family of size continuum that every [distinct] two intersect in a finite set. Can we do that with $\mathcal P(\mathbb R)$, that is a family of size $2^{\frak c}$ many subsets of real numbers that the intersection of any [distinct] two is finite, or at least less than $\frak c$?

The question, if so, asks about $(2^{\frak c})^+$-c.c. in the Boolean algebras $\mathcal B_\kappa=\mathcal P(\mathbb R)/\sim_\kappa$ where $\sim_\kappa$ is the equivalence relation defined as $A\sim B\iff |A\triangle B|<\kappa$. The first question asks for $\mathcal B_\omega$ and the latter asks for $\mathcal B_\mathfrak c$.

Assuming GCH (or at least that $2^{\frak c}=\aleph_2$) gives a relatively simple positive answer to the latter question:

Consider the tree $2^{<\omega_1}$, it is of size $\aleph_1$ so we can encode the nodes as a real numbers. This tree has $2^{\omega_1}=\aleph_2$ many branches, each defines a subset of $\mathbb R$ using the encoding, and every distinct two branches meet at most at countable set of points.

I consulted with several other folks from the department and I was told that most of these questions are very well known, so an answer about consistency and provability is almost certainly out there. Naive Google search got me nowhere, so I came to ask here the following:

  1. In the particular case of the question above, can we say anything in ZFC about the chain-condition of $\mathcal B_\kappa$ for $\omega\leq\kappa\leq\frak c$?

  2. My partial answer above shows that with GCH we have an answer for $\cal B_\frak c$, but does that also answer $\cal B_\omega$ or do we need to assume stronger principles as $\lozenge$ for suitable cardinals?

  3. How far does this generalized, when replacing $2^\omega$ by any infinite cardinal $\mu$, and asking the similar question about $(2^\mu)^+$-c.c. in the similar quotients?

I'd be glad to have a reference to a survey of such results, if it exists.

share|cite|improve this question
There is a nice but old survey by Milner and Prikry in Surveys in combinatorics 1987 LMS Lecture Notes 123 - – François G. Dorais Jun 8 '12 at 14:18
up vote 6 down vote accepted

The answer to your first question (with finite intersections) is negative.

Indeed, if $X$ is an infinite set and $I$ has cardinality greater than that of $X^{\aleph_0}$ then $X$ can't contain $I$ distinct subsets with pairwise finite intersection. This answers your question since $c^{\aleph_0}=c$.

Indeed, let $(A_i)_{i\in I}$ be a family of subsets of $X$ with pairwise finite intersection. Let $B_i$ be the set of infinite countable subsets of $X$ contained in $A_i$. Then the $B_i$ are pairwise disjoint. Moreover, $B_i$ is empty only when $A_i$ is finite, and we can remove such exceptional $i$'s because the number of finite subsets of $X$ is only the cardinality of $X$.

The $B_i$ live in the set of infinite countable subsets of $X$, which has cardinality $X^{\aleph_0}$. So $I$ is at most the cardinal of $X^{\aleph_0}$.

Edit: the obvious generalization of the argument is the following: if $\alpha,\beta,\gamma$ are infinite cardinals, and if $\alpha$ admits $\beta$ subsets with pairwise intersection of cardinal $<\gamma$, then $\beta\le\alpha^\gamma$. In particular, if $\alpha=2^\delta$ and $\gamma\le\delta$ then $\alpha^\gamma=\alpha$, so the conclusion reads as: $2^\delta$ does not admit more that $2^\delta$ subsets with pairwise intersection of cardinal $<\delta$.

share|cite|improve this answer
So this argument can be extended to "countable intersection" if $X$ is large enough, right? – Asaf Karagila Jun 8 '12 at 17:43
@Asaf: you mean, if $|X|^{\aleph_1}=|X|$ where $|X|$ is the cardinal of $X$; this is true if $|X|=2^\alpha$ with $\alpha\ge\aleph_1$ but this does not mean it is true for every cardinal large enough (at least I don't claim it). – YCor Jun 8 '12 at 21:53

It is consistent with ZFC that a set of size $\aleph_1$ does not have $2^{\aleph_1}$ subsets, each of size $\aleph_1$, with all pairwise intersections countable. This is an old result of Jim Baumgartner; see "Almost-disjoint sets, the dense-set problem, and the partition calculus", Ann. Math. Logic 9(1976), 401-439, particularly Theorem 5.6(d) and the remark on page 422 after it. [Caution: I can't check the paper itself now; I'm going by an old e-mail from Jim.]

share|cite|improve this answer

I'll add in that Shelah has used pcf theory to investigate related questions. Typically these results are tucked away inside long papers dealing with other questions, but I know that the last section of [Sh:410] explicitly deals with ``strongly almost disjoint families", and characterizes their existence in terms of pcf.

For example, if $\aleph_0<\kappa\leq\kappa^{\aleph_0}<\lambda$, then the existence of a family of $\lambda^+$ sets in $[\lambda]^{\kappa}$ with pairwise finite intersection is equivalent to a ``pcf statement''.

I'm not sure which version of the paper to link to, as the published version has been reworked a few times. I THINK that the most recent version is here:


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.