3

The Cartesian product of a family $(A_i\mid i\in I)$ is defined as $$\prod\limits_{i\in I}A_i=\{f:I\to\bigcup A_i\mid f(i)\in A_i \text{ for all } i \in I\}$$

Let $(A_i \mid i \in I)$ be a family of non-empty indexed sets where $I$ is infinite, $A_k$ is countably infinite, and $A_i$ is countable for all $i \neq k$. Is $\prod\limits_{i\in I}A_i$ countable?

I found that it's not too hard to conclude when $I$ is finite or when $A_k$ is uncountable. Please give me some hints in this case!

Akira
  • 17,367
  • I think when you cartesian product two sets of countably infinite sets, then the cartesian product is countable infinite too, but I have no proof. – PackSciences Aug 31 '18 at 12:47
  • Assuming the axiom of choice, we have that $\bigcup_{i\in I}$ is uncountable iff $I$ is uncountable, so I think the answer clearly depends on the cardinality of $I$. – blub Aug 31 '18 at 12:48
  • @PackSciences it's a countable union of countable sets –  Aug 31 '18 at 12:49
  • @ChrisCuster did I say it wrong? I reused the term "cartesian product" of the question. – PackSciences Aug 31 '18 at 12:51
  • @PackSciences No. I was just suggesting how it can be proved. –  Aug 31 '18 at 12:52

3 Answers3

1

The Cartesian product will be countable if any of the sets $A_i$ is empty, since then the Cartesian product will also be empty. Let us assume that for all $i$, $A_i\neq \varnothing$.

If there exists a finite subset $J$ of $I$ such that $A_i$ is a singleton for all $i\in I\setminus J$, then the Cartesian product will be countable. To see this, we can rearrange the terms to assume that $A_1, \ldots, A_n$ have more than one element, but $A_{n+1}, A_{n+2}, \ldots$ each have exactly one element. Then there is a bijection between $\prod_{i=1}^\infty A_i$ and $\prod_{i=1}^n A_i$.

Now let us consider the case where infinitely many of the $A_i$ have at least two elements (still assuming each $A_i$ is non-empty). Then $\prod_{i=1}^\infty A_i$ will contain a subset which has the same cardinality at $\{0,1\}^\mathbb{N}$. We should be able to answer the question from here. I will leave the details to you.

  • Hi @bang, on the basis of your and Vera's answers. I presented a detailed proof as an answer below so that I can truly understand your ideas. Please have a check on my proof! Thank you so much – Akira Sep 01 '18 at 09:42
1

If $J:=\{i\in I\mid A_i\text{ contains at least two elements }\}$ then the product contains a subset that has the same cardinality as $\{0,1\}^J$ which on its turn has the same cardinality as powerset $\wp(J)$.

This set is not countable if $J$ is infinite, so in that case the product is not countable.

If $J$ is finite and for every $i\in J$ the set $A_i$ is countable then the product is countable.

If for at least one $j\in J$ the set $A_i$ is not countable then the product is not countable.

If $J$ is empty then the product is a singleton.

Vera
  • 3,469
0

On the basis of two answers, I present a proof in detail. The motivation for this re-formalization is that I would like to write down the proof on my own words, so that I can truly understand every tiny detail of it. All credits go to @Vera and @bang.


Lemma 1: The Cartesian product of a finite amount of countable sets is countable. (I presented a proof here)

Lemma 2: $\mathcal{P}(\mathbb N)$ is uncountable.

Let $J_1 := \{ i \in I \mid |A_i| \ge 2\}$, $J_2 :=J \setminus J_1 = \{i \in I \mid |A_i| = 1\}$, and $a_i$ be the element of $A_i$ for all $i \in J_2$. Consider

$$\begin {array}{l|rcl} G : & \prod\limits_{i\in J_1}A_i & \longrightarrow & \prod\limits_{i\in J}A_i \\ & \phi & \longmapsto & G(\phi) \end{array}$$

where $\phi:J_1 \to \bigcup_{i\in J_1}A_i$ such that $\phi(i) \in A_i$ for all $i \in J_1$.

$G(\phi)$ is defined by $$G(\phi)(i)=\begin{cases}\phi(i) & \text{ if } i \in J_1\\ a_i & \text{ if } i \in J_2 \end {cases}$$

a. $G$ is injective

For $\phi_1,\phi_2 \in \prod\limits_{i\in J_1}A_i$ and $G(\phi_1)=G(\phi_2)$. Then ${G(\phi_1)|}_{J_1}={G(\phi_2)|}_{J_1}$ and ${G(\phi_1)|}_{J_2}={G(\phi_2)|}_{J_2}$. Then $\phi_1(i)=\phi_2(i)$ for all $i \in J_1$ and $a_i = a_i$ for all $i \in J_2$. Thus $\phi_1 = \phi_2$. Hence $G$ is injective.

b. $G$ is surjective

For $\psi \in \prod\limits_{i\in J}A_i$, we define $\phi \in \prod\limits_{i\in J_1}A_i$ by $\phi:={\psi}_{|J_1}$. Then $G(\phi)(i)=\phi(i)$ for all $i \in J_1$ and $G(\phi)(i)=a_i$ for all $i \in J_2$. Then $G(\phi)(i)=\psi(i)$ for all $i \in J_1$ and $G(\phi)(i)=\psi(i)$ for all $i \in J_2$ (Since $A_i = \{a_i\}$ for all $i \in J_2$). Thus $G(\phi)=\psi$. Hence $G$ is surjective.

To sum up, $G$ is bijective.

Case 1: $J_1$ is finite

Then $\prod\limits_{i \in J_1}A_i$ is countable by Lemma 1. Furthermore, $G:\prod\limits_{i \in J_1}A_i \to \prod\limits_{i\in J}A_i$ is bijective. Hence $\prod\limits_{i\in J}A_i$ is countable.

Case 2: $J_1$ is infinite

Let $B_i := \{0,1\}$ for all $i \in \mathbb N$.

Since $\aleph \le |J_1|$ and $|B_i| \le |A_i|$ for all $i \in J_1$, $|\mathcal{P}(\mathbb N)|=|\prod\limits_{i \in \mathbb N}B_i| \le |\prod\limits_{i \in J_1}A_i|$. By Lemma 2, $\mathcal{P}(\mathbb N)$ is uncountable. Thus $\prod\limits_{i \in J_1}A_i$ is uncountable. Furthermore, $G:\prod\limits_{i \in J_1}A_i \to \prod\limits_{i\in J}A_i$ is bijective. Hence $\prod\limits_{i\in J}A_i$ is uncountable.

Akira
  • 17,367
  • It looks good. One suggestion I have about presentation would be to change "1. $J_1$ is finite'' to say "Case $1$:$J_1$ is finite'' and a similar thing for the $J_1$ infinite case. I initially thought you were saying at $1.$ that $J_1$ was finite, rather than saying you were going to consider the case that $J_1$ is finite. –  Sep 01 '18 at 13:16
  • Thank you so much @bangs ! I have edited to reflect your suggestion. – Akira Sep 01 '18 at 15:20
  • @bangs I have asked a question at https://math.stackexchange.com/questions/2907637/define-a-countably-infinite-subset-of-x for several days, but have not received any answer. Although I have some questions that are not answered such as this one. I'm very in need of getting an answer for this question. Could you please help me check it out? – Akira Sep 11 '18 at 23:47