# Measure Theory, $\sigma$-algebra Folland Problem 23 [closed]

I'm preparing for my exam. Can anyone help me in this matter, is confusing to me

thank you very much.

## closed as off-topic by user223391, levap, anomaly, Nikunj, Davide GiraudoMay 31 '16 at 21:11

This question appears to be off-topic. The users who voted to close gave this specific reason:

• "This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level." – Community, levap, anomaly, Nikunj, Davide Giraudo
If this question can be reworded to fit the rules in the help center, please edit the question.

• I'm assuming $\mathcal{P}(\mathbb{Q})$ refers to the Borel $\sigma$-algebra on $\mathbb{Q}$? – Jason Jan 30 '15 at 5:01
• Do you know what $\mathcal{P}(\mathbb{Q})$ is? Do you recognise that $\mathbb{Q} \cap (a,b) = \{(a,b) : a < b, a,b \in \mathbb{Q}\}$ and that for any $r \in \mathbb{Q}$, we have $\{r\} = \cap_{n=1}^\infty (r-n^{-1},r]$? Can you prove this cannot be accomplished with finite unions and complementation? – snar Jan 30 '15 at 5:04
• What are you stuck on? All of these can be verified directly from the relevant defintions. – anomaly May 31 '16 at 18:05
• @Jason $\mathcal P(\Bbb Q)$ is the set of all the subsets of $\Bbb Q$. Which is incidentally the Borel $\sigma$-algebra of $\Bbb Q$ as well, but that's kind-of point (b) of the question. – user228113 May 31 '16 at 19:12

Part b:

Here $\mathcal{P}(\mathbb Q)$ is simply the powerset of $\mathbb Q$. Thus we are trying to show that every subset of $\mathbb Q$ can be obtained by countably many union and complement operations on the sets $(a,b]$.

First observe that any singleton $\{r\}$ can be obtained in the manner, since $$\{r\}=\bigcap_{\substack{s<r\\s\in \mathbb Q}}(s,r].$$ Next observe that for any $R\subset \mathbb Q$, we have $$R=\bigcup_{r\in R}\{r\}.$$ Consequently $\mathcal{P}(\mathbb Q)$ is the $\sigma$-algebra generated by the sets $(a,b]$.

Part c:

Define $\mu(\{r\})$ to be any number you like, for each $r\in\mathbb{Q}$ and extend to all of $\mathcal{P}(\mathbb{Q})$ by $\sigma$-additivity. You get $\mu((a,b])=\infty=\mu_0((a,b])$.

Let $\mathcal{A}$ be the collection of finite unions of sets of the form $(a,b]\cap\mathbb{Q}$ where $-\infty \leq a < b\leq \infty$

a.) $\mathcal{A}$ is an $\sigma$-algebra on $\mathbb{Q}$

b.) The $\sigma$-algebra generated by $\mathbb{A}$ is $\mathcal{P}(\mathbb{Q})$

c.) Define $\mu_0$ on $\mathcal{A}$ by $\mu_0(\emptyset) = 0$ and $\mu_0(A) = \infty$. Then $\mu_0$ is a premeasure on $\mathcal{A}$, and there is more than one measure on $\mathcal{P}(\mathbb{Q})$ whose restriction to $\mathcal{A}$ is $\mu_0$.

Proof a.) - Let $\varepsilon = \{(a,b]\cap \mathbb{Q}: a,b\in\overline{\mathbb{R}}\}$. Clearly $\emptyset = (0,0]\cap\mathbb{Q}\in\varepsilon$. If $(a_1,b_1]\cap\mathbb{Q}\in\varepsilon$ and $(a_2,b_2]\cap\mathbb{Q}\in\varepsilon$ then their intersection is $(\max\{a_1,a_2\},\min\{b_1,b_2\})\cap\mathbb{Q}\in \varepsilon$. Moreover, the complement of $(a,b]\cap\mathbb{Q}$ is $(-\infty,a]\cap\mathbb{Q}\cup (b,\infty]\cap\mathbb{Q}$ which is a disjoint union of elements of $\varepsilon$ if $a\leq b$. If $a>b$ then the complement of $(a,b]\cap\mathbb{Q}$ is $(-\infty, \infty]\cap\mathbb{Q}\in\varepsilon$. This shows that $\varepsilon$ is an elementary family of subsets of $\mathbb{Q}$, so the collection of finite disjoint unions of members of $\varepsilon$ is an algebra by proposition 1.7. Let $(a_1,b_1]\cap\mathbb{Q}\in\varepsilon$ and $(a_2,b_2]\cap\mathbb{Q}\in\varepsilon$ be not disjoint. Then, $(a_1,b_1]\cap\mathbb{Q}\cup (a_2,b_2]\cap\mathbb{Q} = ((a_1,b_1]\cup (a_2,b_2])\cap\mathbb{Q}$. So we need to figure out what is $(a_1,b_1]\cup (a_2,b_2]$ which we have 4 different outcomes: \begin{align*} (a_1,b_1]\cup (a_2,b_2] =\\ &1. (a_1,b_2] \ \ \text{if} \ \ a_1\leq a_2\leq b_1 < b_2\\ &2. (a_1,b_1] \ \ \text{if} \ \ a_1\leq a_2< b_2 \leq b_1\\ &3. (a_2,b_2] \ \ \text{if} \ \ a_2\leq a_1\leq b_1 \leq b_2\\ &4. (a_2,b_1] \ \ \text{if} \ \ a_2\leq a_1< b_2 \leq b_1\\ \end{align*} which are all in $\varepsilon$ thus $\mathcal{A}$ is the collection of finite disjoint unions of members of $\varepsilon$, so $\mathcal{A}$ is a $\sigma$-algebra.

Proof b.) - Suppose that $r\in\mathbb{Q}$. Then $$\{r\} = \bigcap_{n\geq 1}((r-1/n,r]\cap\mathbb{Q})\in M(\mathcal{A})$$ If $X\subset \mathbb{Q}$, then $X$ is countable, say $X = \{x_1,\ldots\}$. But $\{x_j\}\in M(\mathcal{A})$ for all $j\geq 1$. So $$X = \bigcup_{j\geq 1}\{x_j\}\in M(\mathcal{A})$$ Therefore $M(\mathcal{A}) = \mathcal{P}(\mathbb{Q})$

Proof c.) - It is clear that $\mu_0$ is a premeasure on $\mathcal{A}$. The induced outer measure is $$\mu^*(E) = \inf\{\sum_{1}^{\infty}\mu_0(E_j):E_j\in\mathcal{A},E\subset \bigcup_{1}^{\infty}E_j\}$$ it follows that from theorem 1.14 that $\mu = \mu^*$ is a measure on $\mathcal{P}(\mathbb{Q})$ which extends $\mu_0$.

If $\nu$ is a counting measure on $\mathcal{P}(\mathbb{Q})$ then $\nu(\emptyset) = 0 = \mu_0(\emptyset)$. And for $a<b$, $E = (a,b]\cap\mathbb{Q}$ has infinite cardinality, so $\nu(E) = \infty = \mu_0(E)$ and therefore $\nu|\varepsilon = \mu_0|\varepsilon$. By finite additivity of $\nu$, it follows that $\nu(E) = \infty = \mu_0(E)$ for all $\emptyset \neq E\in\mathcal{A}$. So that $\nu|\mathcal{A} = \mu_0$. But $\nu\{1\} = 1 < \infty = \mu(\{1\})$, so $\nu$ is a different measure on $\mathcal{P}(\mathbb{Q})$ which extends $\mu_0$.