I'm back with a question really close to this one : Does it always exist an infinite subset of sequences that satisfy this property?

Now I'm asking myself : does it exist an uncountable set that is dominated. The exact question :

Let $S \subset \mathbb{N}^{\mathbb{N}} $ be an uncountable set of sequence. Does it exist $B\subset S$ such that

  • $B$ is uncountable

  • $\forall k \in \mathbb{N},\quad \sup\limits_{u\in B} u_k < +\infty$

(We include the axiom of choice)

And I don't know what kind of tools to use to prove or disprove this property. Thanks

  • $\begingroup$ What is $u_k$ here? $\endgroup$ – Wojowu Mar 5 '15 at 19:04
  • $\begingroup$ @Wojowu $u$ is a sequence, $u=(u_0,u_1,\dots)$ with each $u_k$ a natural number. $\endgroup$ – Andrés E. Caicedo Mar 5 '15 at 19:05
  • $\begingroup$ This is a little confused. You're taking the sup of all the $u_k$ for each fixed $k$. So for example if you take $B$ to be the set of all binary sequences, the sup over all $B$ for a given $k$ will be 1. Is that the question you meant to ask? And even if you mean to take the sup of each sequence $u$, if the sequences are binary all the sups will still be 1. And there are uncountably many binary sequences. $\endgroup$ – user4894 Mar 5 '15 at 19:09
  • 2
    $\begingroup$ @user4894 What do you mean? $S$ is given, so you may not get the chance to pick an uncountable subset of $S$ consisting of binary sequences. The question is whether for each uncountable $S$ there is such uncountable subset $B$. (The problem is that the obvious thinning out process does not work: Given $S$, you can find uncountable subsets $S\supset S_0\supset S_1\supset\dots$ where, for each $k$ there is a constant $n_k$ such that $u_k=n_k$ for all $u\in S_k$. But the intersection of the $S_k$ is either a singleton or empty.) $\endgroup$ – Andrés E. Caicedo Mar 5 '15 at 19:13
  • $\begingroup$ @AndresCaicedo Oh I see. $\endgroup$ – user4894 Mar 5 '15 at 19:14

The answer is "yes" if and only if the bounding number $\mathfrak{b}$ is larger than $\omega_1$. Here, by $\mathfrak{b}$, I mean the minimum cardinality of a set $S\subseteq \mathbb{N}^\mathbb{N}$ which is unbounded in the $<^*$ order, where $$ x <^* y \iff \exists n\in\mathbb{N}\forall m\ge n\;\; x(m) < y(m) $$

$\mathfrak{b}$ can take many values across different models of set theory, so this shows that your question is independent of ZFC.

First, suppose $\mathfrak{b} = \omega_1$. This means there is a set $S$ of sequences in $\mathbb{N}^\mathbb{N}$, of cardinality $\omega_1$, such that for every $s\in \mathbb{N}^\mathbb{N}$, there is some $u\in S$ such that $u <^* s$ does not hold. Since $S$ has cardinality $\omega_1$, and $\mathbb{N}^\mathbb{N}$ is countably-directed under $<^*$, we may construct a sequence $x_\alpha\in\mathbb{N}^\mathbb{N}$ ($\alpha < \omega_1$) such that $x_\alpha <^* x_\beta$ whenever $\alpha < \beta < \omega_1$, and $X = \{x_\alpha\;|\; \alpha < \omega_1\}$ has the same property. Then any uncountable subset of $X$ is also unbounded, and this violates what you want.

On the other hand, suppose $\mathfrak{b} > \omega_1$, and let $S\subseteq\mathbb{N}^\mathbb{N}$ be an uncountable set. Let $T\subseteq S$ be a subset with cardinality $\omega_1$. Since $\mathfrak{b} > \omega_1$, we can find an $x$ such that $t <^* x$ for all $t\in T$. A Pigeonhole argument shows that there exists $k\in\mathbb{N}$ and a finite sequence $\sigma\in\mathbb{N}^k$ such that for all $u$ in some uncountable subset $U$ of $T$, and for all $n\ge k$, $u(n) < x(n)$, and moreover $u$ looks just like $\sigma$ up to $k$. Then $U$ satisfies what you want.

  • $\begingroup$ Thanks you very much. The result is quite interesting $\endgroup$ – Tryss Mar 6 '15 at 21:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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